Patent Application
20060229211
This invention relates to thin films of high temperature superconducting compositions optimized for RF applications and a method for manufacturing them, more specifically rare earth compositions of (RE)Ba2Cu3O7-δ deviating significantly from the 1:2:3 stoichiometry.
Rare earth oxide superconductors and their ability to superconduct at significantly higher temperatures than previously recorded was first reported by J. G. Bednorz and R. A. Muller in 1986 in regard to mixtures of lanthanum, barium, copper and oxygen in an article entitled “Possible High Tc Superconductivity in the Ba—La—Cu—O system.” (64 Z. Phys. B.—Condensed Matter, pp 189-193 (1986)). Bednorz and Muller described Ba—La—Cu—O compositions that offered a substantial increase in the critical temperature at which the material becomes superconducting over what had been previously known for other classes of materials. Here, the composition was La5-xBaxCu5O5(3-y) where x=0.75-1, y>0, and the abrupt change in resistivity occurred in the 30 Kelvin range.
This contribution led to intensive investigation in order to develop materials having even higher transition temperatures, preferably above 77 Kelvin as this enabled the use of liquid nitrogen to cool the superconducting equipment. In 1987, C. W. Chu and co-workers at the University of Houston found that the onset Tc of the La—Ba—Cu—O compound could by increased to over 50 K by the application of pressure. (Phys. Rev. Lett. 58.405 (1987); Science 235, 567 (1987)).
Chu and coworkers at Houston and at the University of Alabama subsequently discovered a mixed-phase Y—Ba—Cu—O system onset having Tc values near 90 K and a zero-resistance state at ˜70 K. This compound had the nominal composition Y1-2Ba0.8CuO4-δ. (Phys. Rev. Lett. 58, 908 (1987). Chu and coworkers as well as scientists at AT&T and IBM later showed this compound to consist of two phases of nominal composition Y2BaCuO5 (the “green” phase) and YBa2Cu3O6+x (the “black” phase). The latter phase was determined to be the superconducting phase, whereas the former was semiconducting (Cava et al., Phys. Rev. Lett. 58, 1676 (1987); Hazen et al., Phys. Rev. B 35, 7238 (1987); Grant et al., Phys. Rev. Lett. 35, 7242 (1987).
Superconductivity near 90 K was also reported in a mixed-phase Lu—Ba—Cu—O compound by Moodenbaugh and coworkers (Phys. Rev. Lett. 58, 1885 (1987). Chu et al. also identified superconductivity above 90 K for compounds of the formula ABa2Cu3O6+x, where A=Y, La, Nd, Sm, Eu, Gd, Ho, Er, or Lu (Phys. Rev. Lett. 58, 1891 (1987).
The data from these differing Rare Earth (RE)BCO (RE=rare earth, B=Ba, C=Cu) compounds demonstrated that for this class of compounds, the superconductivity is associated with the CuO2—Ba—CuO2—Ba—CuO2 plane assembly which can be disrupted by the A cations only along the c-axis.
Following this discovery, research was focused on the YBCO class of compounds with high temperature superconducting (HTS) properties. B. Batlogg first discovered and isolated the single crystallographic phase responsible for the superconducting properties of the YBCO compound. (B. Batlogg, U.S. Pat. No 6,636,603). In isolating this single perovskite phase of a composition, Batlogg admonished that the composition was essential to isolation of the phase and that it must be within 10% of the M2M′Cu3O7-δ composition where M is a divalent cation preferably barium and M′ is a trivalent cation preferably yttrium.
Other studies have investigated both the effects of substitution of various rare earth elements for yttrium and of varying the 1:2:3 ratio of Y:Ba:Cu on the superconducting properties of HTS compositions. Multiple studies have shown the ability to partially or completely substitute rare earth elements except Pr, Ce and Tb and maintain a Tc of approximately 90 K for the resulting (RE)BCO composition. (S. Jin, Physica C 173, pp 75-79 (1991)). Additionally, further studies show that the c-axis coherence length and the Tc value increase with increasing ionic radius of the rare earth element substituted for yttrium (G. V. M. Williams, Physica C 258, pp 41-46 (1996)).
Building on these discoveries, P. Chaudhari and his co-workers at IBM developed a method for making thin films of high temperature superconducting oxides with a nominal composition of (RE)(AE)2Cu3O9-y where RE is a rare earth element, AE is an alkaline earth element and y is sufficient to satisfy valence demands. (Chaudhari, U.S. Pat. No. 5,863,869 (1999)). The rare earth elements used included Y, Sc and La, and AE could also be substituted for by Ba, Ca or Sr. Copper was the preferred transition metal for the oxide due to its high superconducting onset temperature and the smooth, uniform properties of the copper oxide films. Using this growth process, Chaudhari was able to obtain YBCO films with superconducting onset temperatures of about 97 Kelvin that exhibited superconducting behavior from 50 Kelvin to in excess of 77 Kelvin. These films were within 15% of the targeted (RE)(AE)2Cu3O9-y composition, and Chaudhari noted that the exact composition was not necessary in order to observe high temperature superconductivity.
However, in another study of (RE)BCO cation exchange in thin films, J MacManus-Driscoll et al. noted that Tc decreased dramatically for off-composition films with substitutions of rare earth (RE) elements on the Ba site such as RE(Ba2-xREx)Cu3Oy where RE=Er or Dy and x>0.1 (14% deviation) and where RE=Ho and x>0 (any deviation). (J. L MacManus-Driscoll, Physica C 232, pp 288-308 (1994). J. MacManus-Driscoll further reported that the oxygen pressure at which the thin films were grown seemed to have an effect on the structural disordering of the RE and Ba cations as did the rare earth ion size. Small rare earth cations substituting for the larger Ba cations would produce large strains on the lattice and therefore an unstable phase which would not likely occur.
Another study of varying the 1:2:3 stoichiometry of YBCO thin films noted that large excesses of yttrium formed ultra small yttrium precipitates leading to increased surface resistance (Rs) and poor microwave quality but that a slightly enhanced copper and yttrium content lead to minimum surface resistance (E. Waffenschmidt, J. Appl. Phys. 77 (1) pg 438-440). Furthermore, N. G. Chew et al. analyzed the effect of slight changes in composition on YBCO thin film structural and electrical properties and discovered that films grown with a stoichiometry close to 1:2:3 or with excess yttrium are smooth while films with excess barium exhibited surface roughness and growth of a-axis-oriented grains. (N. Chew, Appl. Phys. Lett. 57 (19) pp 2016-2018 (1990). These authors further found that there is a well defined YBCO composition where Tc and Jc are maximized and the c-axis lattice constant, (007) x-ray peak width, and surface roughness are minimized. These quantities were optimized for a Ba/Y ratio of 2.22±0.05 (subsequently suggested to instead be equal to 2) and a Cu/(Y+Ba+Cu) ratio of 0.5. Slight changes in cation ratios away from this optimized composition caused significant degradation in the parameters listed above.
W. Prusseit et al. have created an iso-structural Dy-BCO thin film material with improved properties compared to their YBCO films. By substituting dysprosium for yttrium and growing under identical conditions as YBCO, Prusseit created films that deviated only slightly from the 1:2:3 stoichiometry. Compared to their YBCO films, these materials exhibited better chemical stability and enhanced transition temperatures (by 2-3 K), and they also had a 20% reduction in surface resistance (Rs) at 77 K: ˜250 μΩ vs. ˜300 μΩ at 10 GHz, measured in a microwave cavity (W. Prusseit, Physica C 392-396, pp 1225-1228 (2003)). Hein (High-Temperature Superconductor Thin Films at Microwave Frequencies (Springer Tracts in Modern Physics, 155), Berlin, 1999) and others have measured somewhat lower surface resistance, ˜200 μΩ at 10 GHz and 77 K, in cavity measurements of YBCO thin films.
The compositions of these (RE)BCO compounds may be altered substantially from the nominal 1:2:3 stoichiometry in order to optimize their properties for specific applications. It is the primary object of this invention to provide high temperature superconducting thin films that have the lowest possible RF surface resistance (Rs) values as well as the lowest achievable RF nonlinearities. This often requires fabrication of (RE)BCO films that deviate significantly from the 1:2:3 composition. It is another object of this invention to provide a thin film superconductor that is optimized for RF/microwave applications. It is another object of this invention that the film has a low surface resistance. It is another object of this invention that the film has a highly linear RF/microwave surface reactance. It is another object of this invention that the stoichiometry of the film deviates by at least 10% from the standard 1:2:3 stoichiometry and with full substitution for yttrium by a rare earth element.
The films of this invention are high temperature superconducting (HTS) thin films specifically optimized for microwave and RF applications. The prior art (RE)BCO films exhibiting high temperature superconducting properties were nominally of the composition (RE)xBayCu3O7-δ where RE=a rare earth element, preferably yttrium, x=1, y=2 and 0≦δ≦1. This 1:2:3 stoichiometry has since been the focus of much study including varying the rare earth element, full and partial substitutions for RE, for Ba, and for Cu, oxygen doping, and deviations from the 1:2:3 stoichiometry.
The present invention focuses on RE HTS films specifically optimized for microwave and RF applications. The RF/microwave HTS applications require the HTS thin films to have superior microwave properties, specifically low surface resistance, Rs, and highly linear surface reactance, Xs, i.e. high JIMD. As such, the invention is characterized in terms of its physical composition, surface morphology, superconducting properties, and performance characteristics of microwave circuits made from these films
In particular, this invention focuses on compositions having a significant deviation from the 1:2:3 stoichiometry in order to create the films optimized for microwave/RF applications. These films have a RE:Ba ratio of less than 1.8, which deviates more than 10% from the typical ratio of 2, and preferably less than 1.7. The research has shown that the highest quality factor values, Q, representing the surface resistance of patterned films, peak at a particular Ba:RE ratio for each RE and that these ratios deviate significantly from the 1:2:3 stoichiometry.
Additionally, the performance characteristics of the HTS films naturally affect their efficacy in RF/microwave HTS applications. Specifically desirable are low surface resistance, Rs, (<15 micro-ohms at 1.85 GHz and 77 K) and highly linear surface reactance, Xs, i.e., high JIMD values (>107 A/cm2, preferably >5×107 A/cm2 at 77 K). HTS thin films with such properties permit the fabrication of extremely selective filters (60-dB rejection within 0.2% relative frequency, to 100-dB rejection within 0.02% relative frequency, with extremely low in-band insertion loss (<1-dB, preferably <0.2-dB) in an extremely small size (<10-cm2 filter chips), which can handle the interference power levels experienced at the front end of a cellular telephone base station receiver (−50 dBm to −28 dBm, to as high as −12 dBm to 0 dBm or possibly even higher) without producing undesirable distortion in the passband, particularly intermodulation distortion, and more particularly intermodulation distortion products comparable to background noise levels (−173.8 dBm/Hz). Thus, the films of this invention are also characterized by their optimized microwave and RF properties. These and other objects, features and advantages will be apparent from the following more particular description of the preferred embodiments.
Table I displays the maximum Qu values at ˜1.85 GHz obtained for several of our (RE)BCO thin films measured using a patterned test resonator. The measurements were made at 67 K and 77 K for an input power of −10 dBm. This table also shows the Rs values that we have calculated from these Qu values.
As previously mentioned, this invention relates to high temperature superconducting (HTS) thin films with compositions that are optimized for RF/microwave applications and methods for reliably producing such films. As such, the invention is characterized in terms of its physical composition, surface morphology, superconducting properties, and performance characteristics of microwave circuits made from these films (filters, delay lines, couplers, etc.; particularly bandpass and bandreject filters, more particularly bandpass and bandreject preselector filters for cellular telephone base station receivers). The distinction between HTS (RE)BCO films of the prior art and the (RE)BCO films of this invention is found both in the composition that deviates significantly from the 1:2:3 stoichiometry and the highly optimized RF properties of the new composition.
Definitions
For our purposes, a thin film may be defined as a layer (generally, very thin) of a material that is grown, deposited, or otherwise applied to a suitable supporting substrate. The thickness of this film may range from about one nm (10−9 m) to several microns (>10−6 m) thick. The typical range of thin film thickness for many applications is from 100 nm to 1000 nm.
High temperature superconductors (HTS) encompass a broad class of ceramic materials, typically oxides, more typically copper oxides or cuprates, that have a transition temperature or critical temperature, Tc, below which these materials are superconducting. Above this critical temperature, they generally behave as metallic, or “normal,” conducting materials. HTS materials are further generally characterized as having Tc values above about 30 K. Examples of HTS materials include La2CaCu2O6, Bi2Sr2CaCu2O8, YBa2Cu3O7, Tl2Ba2CaCu2O8, HgBa2CaCu2O7, etc. These materials must have a well-defined crystal structure in order to be superconducting, i.e., they must have a very specific regular and repeated arrangement of their constituent atoms.
The rare earth (RE) elements are the 15 lanthanide elements with atomic numbers 57 through 71 that are in Group IIIA of the Periodic Table: lanthanum, cerium, praseodymium, neodymium, promethium, samarium, europium, gadolinium, terbium, dysprosium, holmium, erbium, thulium, ytterbium, and lutetium. Yttrium (atomic number 39), a Group IIIA transition metal, although not a lanthanide is generally included with the REs as it occurs with them in natural minerals and has similar chemical properties. Commonly included with the REs because of their similar properties are scandium (atomic number 21), also a Group IIIA transition metal, and thorium (atomic number 90), an element in the actinide series of the Periodic Table.
Composition
The most ubiquitous HTS material is YBCO, which consists of an ordered amount and arrangement of Y, Ba, Cu, and O atoms. The fundamental repeated unit of this material's specific atomic arrangement is known as the unit cell, consisting nominally of one Y, two Ba, three Cu, and seven O atoms. The size of this compound's orthorhombic unit cell is about 3.82×3.89×11.68 Angstroms in the a-, b-, and c-axis directions, respectively. The atomic ratios needed to form this compound are described by the chemical formula YBa2Cu3O7-δ, where the oxygen content is variable between 6 and 7 atoms per unit cell, or 0≦δ≦1. For single-phase materials with this composition and having high crystalline quality and purity, the Tc value is determined largely by the value of δ. YBCO is a superconductor for δ<˜0.6, with values of δ near 0 being generally preferred in order to provide the highest Tc values.
YBCO is the most widely studied HTS material, and much is known about how to make it in single phase form, i.e., consisting of solely the composition mentioned above and containing no other phases. However, many other similar compounds can also be fabricated that may have similar or superior superconducting properties, depending on the application. These compounds may have Y:Ba:Cu ratios that are different from 1:2:3, and they may also consist of elements other than Y or Ba. A generalized nomenclature for the makeup of this compound may thus be written as M′xMyCu3O7-δ, where M′ may in general be any essentially trivalent ion or combination of ions, and M may be any essentially divalent ion or combination of ions. The ratios of M′:Cu, M:Cu, and M′:M may also vary substantially from the nominal values of 1:3, 2:3, and 1:2, respectively. While the full range of parameter space has not been explored, it is reasonable to believe that compounds with cation ratios deviating from the nominal by as much as 50% may still be superconductors, e.g. 1:6<M′:Cu<1:2, 1:3<M:Cu<1:1, and 1:4<M′:M<3:4. However, significantly altering the composition from the 1:2:3 stoichiometry does affect the specific properties of the composition including critical current density (Jc), normal-state resistivity (ρ), critical temperature (Tc), and surface resistance (Rs).
In order to provide high Tc values, Ba is generally preferred as the divalent element, or M in the above formula. Full or partial substitutions of many elements for Ba tend to decrease Tc or destroy superconductivity altogether. These elements include Sr, La, Pr and Eu. (Y. Xu, Physica C 341-348, pp 613-4 (2000) and X. S. Wu, Physica C 315, pp 215-222 (1999). Similarly, the Cu atoms may be doped with Co, Zn, Ni, etc., the effect of most of which is to decrease Tc, though the absolute effect (e.g., charge transfer or disruption of superconductivity on the Cu—O planes) depends on whether the Cu(1) or Cu(2) sites are affected. (Y. Xu, Phys. Rev. B Vol 53, No. 22, pp 15245-15253 (1996). Some partial substitutions on the Y sites may have a similar effect, such as Ca, Ce, and Pr (L. Tung, Phys. Rev. B Vol 59, No. 6, pp 4504-4512 (1999) and C. R. Fincher, Phys. Rev. Lett. 67 (20) pp 2902-2905 (1991)). However, there are many known partial or complete substitutions for Y that lead to similarly high or greater Tc values than YBCO. Many of these known substitutions come from the rare earth family of elements. In general, rare earth elements that have a larger ionic radius produce higher Tc values for these (RE)BCO compounds (G. V. M. Williams, Physica C 258, pp 41-46 (1996)).
While it is key to maintain the defining property of superconductivity across the range of compositions available for these related compounds, our research has shown that the compositions may be altered substantially from the nominal 1:2:3 stoichiometry in order to tailor their properties for specific applications. For example, compositions near 1:2:3 may be preferred for multilayer or active device applications for which smooth thin film surfaces are of paramount importance. Conversely, optimization of HTS films for RF applications requires the production of thin films that strike a balance between having the lowest possible RF surface resistance (Rs) values and the lowest RF nonlinearities that are achievable. This in turn often requires the fabrication of (RE)BCO films that deviate significantly from the 1:2:3 composition.
The HTS thin films of this invention are optimized for RF applications, and as such they have the lowest possible RF surface resistance (Rs) values and the lowest possible RF nonlinearities. In order to achieve this optimization, the film compositions have the nominal formula (RE)xBayCu3O7-δ where RE is one of the previously defined rare earth elements, preferably Dy, and where the ratio y:x is preferably between about 1.5-1.8, more preferably between about 1.55-1.75, and most preferably between about 1.6-1.7.
Substrates
The superconducting properties of HTS materials are extremely sensitive to their degree of crystalline perfection. This places severe constraints on the choice of a suitable substrate material on which high-quality HTS films may be grown. Some of these constraints include crystal structure, compatibility with the growth process, chemical compatibility, compatibility with the application, as well as other requirements imposed by nature.
Perhaps the most important requirement is the crystal structure. The substrate must have an appropriate lattice match with the HTS film such that epitaxial growth of the film can occur and a well-oriented film will form. A poor lattice match can lead to dislocations, defects, and misoriented grains in the film. In general, the substrate should be available in single-crystal form in order to meet these requirements.
The substrate must be able to withstand the high processing temperatures during the growth process that are required for the crystallization of the HTS compound. In addition, structural integrity and a reasonable thermal expansion match with the HTS film is required in order to prevent strain and cracking of the film during the cool down cycle from the growth temperature or from any other subsequent thermal cyclings.
The substrate must be chemically compatible with (RE)BCO, non-reactive, and with minimal diffusion into the film at high temperature.
The substrate must be available in a size large enough for the intended use of the HTS thin film. For example, certain passive microwave circuits or high-volume electronics applications require a large substrate size. A minimum substrate of 2″ in diameter is typical for these applications, though larger sizes are often desirable if available. The substrate may also be required to have physical properties that are compatible with experimental measurement techniques or applications. For most applications the substrates should be stable, mechanically robust insulators. Other requirements may include transparency in the infrared for optical transmission measurements, constituent elements or structure that do not interfere with spectroscopic measurements such as Rutherford backscattering (RBS) or energy-dispersive x-ray analysis (EDX), and a low dielectric constant and loss tangent for microwave measurements and applications at the intended temperature of operation.
A handful of single-crystal substrates meet some or all of these requirements. Examples include MgO, Al2O3, LaAlO3, NdGaO3, (La0.18Sr0.82)(Al0.59Ta0.41)O3, and SrTiO3. The last four have an excellent lattice match to (RE)BCO. The high dielectric constant and loss tangent of SrTiO3 make it useless for microwave applications, however. LaAlO3 and NdGaO3 are better in this regard, though LaAlO3 suffers from the fact that it tends to twin, and these twin boundaries can be formed and become mobile at typical processing temperatures. Al2O3 is a low-loss substrate and is widely available in several different orientations and sizes. However, it reacts strongly with (RE)BCO at high temperatures, requiring the use of an appropriate buffer layer. In addition, Al2O3 has a poor thermal expansion match to (RE)BCO, causing a tendency for the films to crack upon cooldown. MgO has relatively low loss and a good thermal expansion match to (RE)BCO, making it a good choice for RF applications. However, MgO has a much larger lattice mismatch than the other examples listed above, so that great care must be taken to insure that the (RE)BCO films grown on MgO are well oriented. In particular, it is relatively common for (RE)BCO films grown on MgO to contain in-plane-rotated grains and 45° grain boundaries. (B. H. Moeckly, Appl. Phys. Lett. 57, 1687-89 (1990). The minimization of the amount of these high-angle grain boundaries is mandatory for good microwave performance, particularly for high RF linearity. Certain MgO substrate surface treatments may be instituted to help control the number of high-angle grain boundaries, but greater effort is required to further suppress formation of these grain boundaries, particularly for demanding RF applications. The growth method, growth conditions, and particularly the composition of the (RE)BCO films must all be chosen and adjusted to minimize the amount of 45° grains in films grown on MgO.
Film Morphology and Microstructure
The anisotropic transport properties of (RE)BCO, its orthorhombic crystal structure, and its small superconductive coherence length mean that the (RE)BCO films must have excellent crystalline structure and orientation. This is particularly true in order to obtain good microwave properties. Hence, the films must be substantially free of secondary phases, they must possess good epitaxy both in-plane (parallel to the substrate surface) and out-of-plane (perpendicular to the substrate surface). Typically, the c-axis of (RE)BCO is aligned perpendicular to the substrate surface. All the grains in the film must be so aligned, and they must be highly aligned with respect to one another. The degree of this crystalline order is typically characterized by θ-2θ x-ray diffraction scans, where the requirements are the existence of only c-axis-oriented (001) spectral lines having narrow peak widths, and also narrow peak widths of the so-called ω-scan, or rocking curve scan about a given Bragg angle. The θ-2θ measurement can also detect the presence of spectral lines due to a-axis-oriented grains within the film. These grains may also be detected by a χ-scan about an appropriate Bragg angle.
For a thin film with good microwave properties, the amount of a-axis grains in the film is ideally zero, so that the intensity of a-axis x-ray peaks relative to c-axis x-ray peaks for a c-axis-oriented film is ideally zero, and preferably much less than 1%. In addition, the c-axis-oriented grains should also be in-plane oriented, meaning that they are in registry with each other and with the substrate crystal structure. Grains that are rotated with respect to the overall in-plane lattice structure lead to nonzero-degree angle grain boundaries. The superconducting transport across such nonzero-angle grain boundaries, in particular high-angle grain boundaries and 45° grain boundaries, is degraded likely due to strain, the high oxygen mobility, and small coherence length of (RE)BCO (B. H. Moeckly et al., Phys. Rev. B 47, 400 (1993). Jc, Rs, and the RF nonlinearities may all be adversely affected by the presence of these high-angle grain boundaries. The presence of these rotated grains and grain boundaries may be detected by φ-scan x-ray measurements taken about an appropriate Bragg angle. Ideally, the amount of nonaligned φ-scan peaks should be zero, and preferably less than 0.1% of the magnitude of the aligned peaks, more preferably less than 0.05%, and most preferably less than 0.02%.
The surface morphology of (RE)BCO thin films is typically measured by scanning probe profilometry, atomic force microscopy (AFM), and scanning electron microscopy (SEM). In general, smooth films are preferred for applications, though some degree of surface roughness may be tolerated in deference to the optimization of other important properties such as Jc and Rs. Still, it is desirable to have an RMS surface roughness as determined by AFM, say, which is less than ˜10 nm.
Film Characterization Methodology
The (RE)BCO films are further characterized by measuring their composition and their electrical properties, including the dc resistivity (p) as a function of temperature [ρ(T)], Tc value and transition width, critical current density (Jc), and RF surface resistance (Rs). The films are also subsequently patterned into RF circuits for which we measure the unloaded quality factor (Q) values, intermodulation distortion (IMD), and nonlinear critical current density (JIMD).
The composition of the films of this invention was measured using Rutherford backscattering spectrometry (RBS) and inductively coupled plasma spectroscopy (ICP). These techniques are both capable of a high degree of accuracy and precision, though achieving a measurement accuracy of 1σ or 2σ equal to 1% is a difficult task and requires more care than is the norm for these techniques. In the RBS analysis technique, fast, light ions (typically He ions or alpha particles) are accelerated toward the sample; some of these ions are backscattered due to Rutherford (Coulomb) scattering from atomic nuclei within the sample, and the energy spectrum of those backscattered particles is analyzed. The ion energies are typically in the range of several hundred to several thousand keV, and the energy of a backscattered ion depends on the mass of the target atom with which it has collided. Thus, the energy spectrum of the backscattered ions allows identification of the elements comprising the sample and their ratios (stoichiometry). In addition, as the incident ions traverse the sample, they lose energy due to inelastic scattering with electrons. This energy loss occurs in a known way and therefore allows determination of sample composition as a function of depth. However, for thick films, the spectral peaks of the measured constituent elements can overlap, requiring careful fitting of the spectra to extract the composition, and this procedure involves uncertainty and can introduce error. Therefore, in order to obtain the highest accuracy by simply counting the number of counts under each peak, sufficiently thin films must be used so that the peaks due to RE, Ba, and Cu can be completely separated. We have grown sufficiently thin (RE)BCO films for this purpose, and the results of these measurements have shown a compositional accuracy of 2σ≦±1%. Note that this measurement technique is quantitative and does not require the use of a comparison standard.
In the ICP technique, the thin films are digested in an acidic solution which is then introduced into a high-temperature (up to 10,000° C.) plasma discharge. The plasma ionizes and excites the constituent atoms in the solution, and as these atoms decay to a lower energy state, they emit light of a characteristic wavelength that can be detected by a high-resolution spectrometer. This is the so-called ICP-AES (atomic emission or optical emission spectroscopy) technique. ICP hence permits measurement of multiple elements simultaneously. ICP-AES has detection limits typically at the μg/L level in aqueous solutions. This technique can be very accurate and precise; an accuracy of 1σ<±1% is obtainable with careful measurement. The method requires the use of a comparison standard. It does not have an accuracy limitation as a function of thin film thickness, however, as does RBS. Hence in testing our compositions, we have used RBS and ICP together. First, we have made careful RBS measurements on very thin films in order to determine their composition to a high degree of accuracy. We have then confirmed that the ICP measurements on these same samples agree with the RBS numbers. This allows us to have confidence that the ICP-AES measurement of thicker (RE)BCO films shares this same degree of desired accuracy, i.e., 1σ<±1%.
The dc resistivity p is measured by a standard four-point-probe technique. The room-temperature resistivity of high-quality (RE)BCO films is typically between 150 and 300 μΩcm, though this value varies as a function of RE element and of film composition.
The Tc values of the (RE)BCO samples prepared by the process of this invention are 88.5(5), 88.9(5), 89.2(5), 89.6(5), and 94.5(8) K for Er, Y, Ho, Dy, and Nd, respectively. These values were measured immediately following deposition. Since the films are oxygen overdoped as judged by the slope of the R-T curves, the measured Tc values are slightly lower than the highest values known for these compounds.
The RF surface resistance of (RE)BCO thin films may be measured in a number of ways, including cavity or parallel plate resonator techniques using bulk (unpatterned) films. Rs is typically measured at frequencies between a few hundred MHz and 10s of GHz. Rs may also be extracted from the Q measurements of patterned resonators of various kinds, e.g., microstrip, quasi-lumped element, etc. Extraction of Rs from the measured Q values of these structures requires careful modeling of the resonator performance to determine the geometric parameter ΓQ. The relationship between Rs and Q can be written as
where w0 is the resonant frequency, ΓQ is a parameter that depends only on the resonator geometry, and Q is the measured unloaded quality factor of the resonator. The extracted Rs value of patterned structures is typically higher than the Rs value obtained by direct measurement of the bulk films in an RF cavity. This may be caused by patterning the film, which may introduce defects that can add additional resistive RF losses in the Q measurement. It may also arise from uncertainties in ΓQ or the non-uniformity of the current density in microstrip resonators which is generally not present in bulk film measurement systems.
Device Performance Characterization
For evaluation of the RF properties of our (RE)BCO films and for determining the utility of these materials for microwave filter applications, we have fabricated microwave resonators and filters from these films. These passive devices require a ground plane and hence necessitate depositing double-sided films. Quasi-lumped element resonators were patterned using standard photolithographic processing and inert ion etching. The geometry of our test resonator is shown in
The Q values of (RE)BCO filters can degrade as a function of increasing input power. The ability of (RE)BCO filters to maintain high Q values as a function of increasing input power is an important requirement for high performance filter systems.
The input power levels to the (RE)BCO filter also affects their performance by generating different amounts of intermodulation distortion, as described below.
We further evaluated these materials by growing thin films on 2″ MgO substrates and patterning them into 10-pole filter circuits of a type suitable for commercial cellular communications applications.
We fabricated B-band cellular microwave filters from several (RE)BCO thin films which were grown by in situ reactive coevaporation onto 2″ MgO substrates. Each double-sided wafer yields two filters, each having a size of 18 mm×34 mm. The patterned (RE)BCO structures are quasi-elliptic 10-pole filters with 3 pairs of transmission zeros on either side of the frequency passband.
Intermodulation distortion in HTS filters arises due to nonlinearity of the microwave surface reactance, X3, of the thin films. (R. B. Hammond et al, J. Appl. Phys. 84 (10) pp 5662-5667 (1998)). In general, at high microwave current densities in HTS thin films Xs ceases to be constant and independent of current density, and begins to increase with increasing current density. Commonly there is a maximum current density, JIMD, at which Xs retains its low current density value, and above which Xs increases. In this paper by Hammond et al, the relationships between measured parameters and the material parameter JIMD are described. This relationship can be summarized as follows
here QL is the loaded quality factor of the resonator, ω0 is the resonant frequency, these two functions depend on the filter function to be realized, ΓIMD is a factor which depends only on the geometry of the resonator, and PIN and POUT are the input and output powers from an intermodulation measurement.
The out-of-band IMD test requirement corresponds to a minimum JIMD in the HTS thin film of 1×107 A/cm2. The DBCO films surpass the specification by 14 dB, which here corresponds to a factor of 5. Thus, the DBCO films have a JIMD of 5×107 A/cm2. For filter applications JIMD in HTS thin films must be >1×107 A/cm2, more preferably >2×107 A/cm2, and most preferably >3×107 A/cm2.
Methods of Manufacture
We have grown our (RE)BCO thin films using an in situ reactive coevaporation (RCE) deposition technique which has been successfully used to manufacture large-area YBCO HTS thin films. This is a fabrication technique that readily lends itself to high volume film production and manufacturability. The yield of high-performance microwave filters made from films grown by RCE is typically >90%. A key component of this growth method is the use of a radiative heater that internally maintains an oxygen partial pressure that is greater than ˜10 mTorr. The heater also incorporates a window that allows exposure of the rotating substrates to high vacuum, where evaporation and deposition of the source materials occurs. Our substrates are typically MgO single crystals up to 2″ in diameter that are rotated continuously between the window and the oxidation pocket at 300 rpm. The chamber ambient pressure away from the pocket is ˜10−5 Torr. This configuration provides sufficient oxygen pressure for stability of the high-Tc phase while the metallic evaporation sources are simultaneously free from oxidation, and the evaporated species are free from scattering. The rare earth elements Er, Ho, and Dy are evaporated from electron beam sources, Nd and Cu are evaporated from either electron beam sources or resistive sources, and Ba is evaporated from a thermal furnace or a resistive source. The typical deposition rate is ˜2.5 Å/sec. The deposition temperature for the films discussed here is 760 to 790° C., and the film thickness is about 700 nm. The films were deposited directly onto MgO substrates, with the exception of Nd-BCO, which presently requires a thin buffer layer in order to achieve the best results.
Unlike yttrium, which melts readily, some rare earth elements such as Er, Ho, and Dy sublime during e-beam evaporation, thereby making compositional control more challenging. We routinely use quartz crystal monitors (QCM) as our primary rate controllers. However, the subliming materials are never molten at our evaporation rates; rather, the electron beam digs a hole in the metallic source material so that the plume shape changes significantly during the course of the deposition run. Therefore, the QCMs are not able to correctly monitor the changing amount of RE vapor flux. To alleviate this difficulty we have employed hollow-cathode-lamp (HCL) atomic absorption (AA) evaporation flux sensors to monitor and control these subliming materials. Since the AA light beam passes through the entire plume of evaporated species, this technique can more accurately monitor the amount of evaporated flux.
The oxygen pocket pressure and deposition rate used to achieve optimal results are similar for the (RE)BCO films that we have studied. We have found that the best substrate temperatures for Er, Ho, Dy, and Nd are 780, 790, 790, and 780° C., respectively. These temperatures are significantly higher than the temperature of 760° C. we use to achieve optimal RF properties for YBCO. The use of different growth conditions for the (RE)BCO materials compared to YBCO is mandatory in order to achieve the very best RF properties. For example, higher growth temperatures for the (RE)BCO materials as compared to YBCO are generally required in order to insure the absence of deleterious misaligned grains. The composition must also be optimized for this purpose, as we have discussed. In general, many aspects of film growth affect the defect structure in (RE)BCO thin films, and thus RF properties, including a) growth temperature, b) growth rate, c) oxygen pressure, and d) stoichiometry. Specific choices for (a), (b), and (c) may yield different optimized properties and different optimized compositions.
Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity and understanding, it may be readily apparent to those of ordinary skill in the art in light of the teachings of this invention that certain changes and modifications may be made thereto without departing from the spirit or scope of the appended claims.
This application claims priority pursuant to 35 U.S.C. § 119(e) to U.S. Provisional Application No. 60/639,043, filed Dec. 23, 2004.
| Number | Date | Country | |
|---|---|---|---|
| 60639043 | Dec 2004 | US |
